Kinetic investigation of the cyclopropanation process of fullerene C60 by halogenmethyl ketones under the conditions of the Bingel reaction

Article abstract of DOI:10.1039/c9nj06326a

The kinetics of the Bingel reaction with halogenmethyl ketones and C₆₀ fullerene has been studied in streaming mode by sampling the reaction mixture at different time intervals and separating the components using HPLC. A quantum-chemical simulation of this cyclopropanation process has been carried out with the DFT method. The activation parameters of the cycloaddition process were determined theoretically and experimentally and correlated with each other. It has been revealed that the use of chloromethyl ketone as the cyclopropanation agent is preferable to its brominated analogue, and a two-fold excess of the substrate with respect to fullerene is the best option for the selective synthesis of mono-adducts.

Full text of DOI:10.1039/c9nj06326a

NJC
View Article Online
View Journal
PAPER
Kinetic investigation of the cyclopropanation
process of fullerene C60 by halogenmethyl
ketones under the conditions of the Bingel
reaction†
Cite this:DOI: 10.1039/c9nj06326a
ab
a
Yulya N. Biglova,
*
Ilshat M. Sakhautdinov,
Rustem N. Garifullin,c
Gulnaz F. Sakhautdinovaa and Akhat G. Mustafina
The kinetics of the Bingel reaction with halogenmethyl ketones and C60 fullerene has been studied in
streaming mode by sampling the reaction mixture at different time intervals and separating the
components using HPLC. A quantum-chemical simulation of this cyclopropanation process has been
carried out with the DFT method. The activation parameters of the cycloaddition process were
determined theoretically and experimentally and correlated with each other. It has been revealed that
the use of chloromethyl ketone as the cyclopropanation agent is preferable to its brominated analogue,
and a two-fold excess of the substrate with respect to fullerene is the best option for the selective
synthesis of mono-adducts.
Received 21st December 2019,
Accepted 4th March 2020
DOI: 10.1039/c9nj06326a
rsc.li/njc
control the reaction by physicochemical methods has both
fundamental and applied value.
Introduction
The recently discovered Bingel reaction of nucleophilic fullerene
As a rule, fundamental questions are addressed based on the
cyclopropanation with stabilized carbanions has become the kinetic study of the reaction transformation of fullerene -
1
basic process of the functionalization of the C60 nucleus. This methanofullerene. However, the scientific literature lacks full
reaction is promising for fine organic synthesis because it systematic studies of both the theoretical and experimental nature
demonstrates almost inexhaustible possibilities for obtaining of the reaction, which would enable control of the functionalization
practically significant substances. Due to its versatility (avail- of C60 with high yields of target products with given structures.
ability and variety of initial substrates, good yields of target However, among the few experimental fragmentary studies in the
products and rather mild synthesis conditions), the significance area, it is worth highlighting the publications of Oshima’s group
of this process in methanofullerene production is difficult to devoted to the kinetics of the [2+4], [2+2] and [2+1]-cycloaddition
estimate.2
reactions of various substrates with C60.3–7 The process was mon-
The C60 molecule is highly symmetric, and all its atoms as itored by high-pressure liquid chromatography (HPLC). Fullerene
well as its double and single bonds are equivalent; therefore, consumption and mono-adduct accumulation were simultaneously
the opening of any bond of the framework leads to the same monitored, and the authors confirmed that C60 functionalization is
product. A variety of [2+n]-cycloaddition reactions are promising a first order reaction, whereas the overall reaction is second order.8
in the fullerene functionalization sphere; as a rule, the cases of
Another group of Japanese scientists led by Joshua P. Barham
n = 1, 2, 3 and 4 are studied. The [2+1]-variant is the most proposed a microwave flow reactor, which sustained high tempera-
common and advanced among synthetic organics from all ture while employing short residence times; these reaction
points of view. Because fullerene functionalization is accompa- conditions uniquely allowed the selective synthesis of fullerene/
nied by mono- and polyadduct formation, the search for ways to indene mono-adducts.9 The design of experiment analysis
revealed that the residence time is the most crucial factor for
conversion and selectivity control.
a Ufa Institute of Chemistry, Ufa Federal Research Centre, Russian Academy of
Sciences, 450054 Ufa, Russian Federation. E-mail: bn.yulya@mail.ru
b Department of Chemistry, Bashkir State University, 450076 Ufa,
Russian Federation
Reaction rate constants were determined for the [2+4]-
cycloaddition of 1,3-cyclohexadiene to fullerene.3,5,10 In contrast to
the described method of Diels–Alder fullerene derivative synthesis
with 1,3-cyclohexadiene, the alternative Bingel method involves
attack on the C60 molecule by stabilized carbanions occurring via
the [6,6] bond only.
c Institute of Mathematics, Ufa Federal Research Centre, Russian Academy of
Sciences, 450008 Ufa, Russian Federation
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nj06326a
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
Paper
NJC
The kinetic studies of Bingel [2+1]-cycloaddition with malo- D,L-b-phenyl-a-alanine, nitrosomethylurea, MgSO4, and NaHCO3
nic ester have been presented in the literature.11 The authors (Aldrich) were used without any purification.
performed experiments in both batch and streaming modes by
Acetone, ethyl acetate, methylene chloride and chloroform
sampling the reaction mixture at certain intervals and separating were distilled over P2O5. Toluene and petroleum ether were
the components using HPLC. At the same time, the fullerene boiled and distilled with sodium.
consumption and the mono- and polyaddition product accu-
The acetonitrile and toluene used were HPLC grade, and
mulation were considered. Three kinetic stages of the process they were prepared by filtration through a 0.45 mm membrane
were simulated: the first is the formation of a reactive inter- filter.
mediate compound, diethyl bromomalonate; the second is the
All physicochemical constants of the solvents corresponded
interaction of diethyl bromomalonate and C60, resulting in a to the reference data.
mono-adduct; and the third is the interaction of a mono-adduct
and diethyl bromomalonate to form a bis-adduct.
The HBr and HCl qualifications of ‘‘extra pure’’ were used
for 48% and 37% solutions, respectively.
Kinetic equations were used to determine the k1, k2 and k3
rate constants for each interaction step by selecting experimental
Instrumental methods
data from different mixing modes. Based on the experimental IR spectra were recorded on an IR-Prestige-21 instrument (Fourier
data, the rate constants of the indicated stages were calculated transform spectrophotometer – Shimadzu) in thin films or in
using the least squares method. However, the scientists did not Vaseline oil. NMR spectra were obtained on a Bruker-AM 500
determine the activation process parameters for the complete spectrometer with working frequencies of 500.13 MHz (1H) and
reaction of [2+1]-cycloaddition to fullerene.
125.76 MHz (13C), using tetramethylsilane as the internal standard.
As for the quantum-chemical study of methanofullerene To ensure correct assignment of signals in the NMR spectra, homo-
synthesis, the influence of 12 substituents in diphenyldiazo- and heteronuclear two-dimensional correlation methods (COSY,
methane on the reactivity of fullerenes was studied in a report12 NOESY, HSQC and HMBC) were used to study the reaction pro-
using density functional theory. The substituent effect changes ducts. The UV-Vis-NIR absorption spectral measurements were
the reactivity of the molecule, and its study is crucial for carried out with a Shimadzu UV-1800 spectrometer. Mass spectra
determining the organic reaction mechanism of C60. The purpose were obtained on a Shimadzu LCMS-2010EV chromato-mass
of the substitution is to determine how it affects the reaction spectrometer in chemical ionization mode at atmospheric pressure,
equilibrium or the free activation energy and, therefore, the and MALDI mass spectra were recorded on an ULTRAFLEX III mass
structure of the transition complex. Thermochemical data reveal spectrometer (Bruker Daltonik GmbH, Bremen, Germany). Melting
that electron-donating groups often increase the DG value of the points were determined on a Boetius heating stage. The course of
activation barrier, whereas electron-withdrawing groups decrease each reaction was monitored using thin-layer chromatography
it. In all cases, solvents such as toluene and acetonitrile only (TLC) on Sorbfil PTSKh-AF-A plates. Compounds were visualized
slightly affect the energy parameters of the process.
using iodine vapors and by spraying the plates with a ninhydrin
Theoretical studies of cycloaddition of bromomalonates to developing solution or with an anisic aldehyde solution fol-
higher endohedral metallofullerenes are also known in the lowed by heating at 100 1C to 120 1C. Reaction products were
scientific literature.13–18 It is shown that the process is carried isolated by column chromatography over Chemapol silica gel
out by a two-step mechanism: initially, bromomalonate reacts (40/100 and 100/160 mm).
with the C60 nucleus by nucleophilic addition, with the formation
of a negatively charged intermediate. At the second stage, intra-
Compounding procedures
molecular closure of the cyclopropane ring takes place during N-Phthalyl-b-phenyl-a-alanine (I) was synthesized by the method
bromide displacement with a carbanion to obtain the final mono- described in ref. 20, and its physicochemical data corresponded
adduct, and the final stage is limiting.19
to the literature.21
As can be seen, complete kinetic studies of the Bingel
Diazo compound (III) was prepared under Arndt–Eistert
cyclopropanation process of C60 combining quantum chemical reaction conditions. Freshly distilled thionyl chloride (2.7 mmol)
and experimental approaches are lacking. was added to D,L-N-phthalyl-b-phenyl-a-alanine (I) suspension 1
Consequently, the goal of this work is to perform an overall in 5 ml dry benzene. A flask provided with a reflux condenser
kinetic study of C60 cyclopropanation with halogenmethyl was fitted with a calcium chloride tube (for exclusion of humid-
ketones by the Bingel reaction, revealing its laws and features. The ity), and the mixture was refluxed for 2.5 h. The solvent and
result is generalized and can be useful for other cyclopropanation excess thionyl chloride were removed by distillation, and the
agents in the synthesis of various methanofullerenes.
resulting acid chloride was used without further purification.
Then, a solution of acid chloride (II) in 10 ml of methylene
chloride was added dropwise with cooling (0 1C) and stirring to
an ethereal solution of diazomethane obtained from 2 mmol
nitrosomethylurea. When the gas evolution was complete, the
reaction mixture was stirred for additional 0.5 h, and the solvent
Experimental
Reagents, solvents and materials
Fullerene C60 (99.5% of the main substance), 1,8-diazabicyclo- was evaporated. The residue was chromatographed on silica gel
[5.4.0]undec-7-ene (DBU), phthalic anhydride, thionyl chloride, (petroleum ether/ethyl acetate, 8 : 2, v/v).
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
NJC
Paper
2-(1-Benzyl-3-diazo-2-oxopropyl)-1H-indene-1,3(2H)-dione (III), 142.74, 142.88, 142.95, 143.01, 143.13, 143.32, 143.34, 143.62,
78% yield.22 1H NMR spectrum (CDCl3, d ppm, J Hz): 3.52 (m, 2H, 143.67, 143.96, 144.02, 144.41, 144.45, 144.54, 144.62, 144.69,
CH2), 5.08 (dd, 1H, CH, 3J 5.9, 3J 10.5), 5.39 (s, 1H, CH), 7.12 (m, 144.82, 145.01, 145.05, 145.13, 145.16, 145.20, 145.27, 145.29,
5H, C6H5), 7.67–7.79 (dm, 4H, C6H4). 13C NMR spectrum: 33.94, 145.36, 145.48, 145.65, 145.72, 147.55, 147.68, 167.79, 195.94.
54.28, 58.35, 123.55, 127.06, 128.61, 128.95, 131.39, 134.46, UV-Vis (CH2Cl2, nm): 259, 327, 427, 494, 689. FT-IR (m.o.): 2916,
136.68, 167.59, 189.31. Found, %: C 67.68, H 4.11, N 14.01. 2849, 1775, 1714, 1382, 1265, 738, 704. C78H13NO3 MS (MALDI-
C18H13N3O3. Calculated, %: C 67.71, H 4.10, N 13.69.
TOF), m/z: 1011.070 (M+) (calc., m/z: 1011.089).
Chloroketones and bromoketones (Cl-K and Br-K) were
2,20-[Bis-cyclopropano[60]fullerene bis(1-oxo-3-phenylpropane-
obtained from the diazo compound (III) by the following 1,2-diyl)]bis(1H-isoindole-1,3(2H)-dione) (DMF), brown powder,
method:
yield 22%. 1H NMR spectrum (CDCl3, d ppm, J Hz): 3.78
To 1 mmol diazoketone (III) solution in 10 ml of methylene (m, 2Ha, 2CH2), 3.91 (m, 2Hb, 2CH2), 4.96 (m, 2H, 2CH),
chloride, 1 ml of a 37% HCl solution or a 48% HBr solution was 5.59–5.83 (m, 1H, CH), 7.32 (m, 10H, 2C6H5), 7.72 and 7.81
added with stirring. After gas evolution was complete, the (dm, 4H, C6H4), 7.75 and 7.92 (dm, 4H, C6H4). 13C NMR
solution was stirred for 1 h. Then, the organic layer was separated spectrum: 34.03, 34.23, 42.25, 42.34, 42.45, 44.42, 44.50, 57.73,
and washed with 3 ml of a 5% water solution of NaHCO3. The 60.77, 60.92, 70.91, 71.65, 71.80, 71.90, 123.70, 123.84, 123.89,
organic layer was dried over MgSO4, the solvent was evaporated, 124.02, 125.34, 27.08, 127.19, 128.27, 128.66, 128.85, 128.99,
and the residue was chromatographed on silica gel (petroleum 129.08, 131.38, 131.47, 131.52, 134.45, 134.55, 134.62, 136.19,
ether/ethyl acetate, 8 : 2, v/v).
136.39, 137.88, 139.05, 139.15, 139.35, 139.41, 139.83, 139.98,
2-(4-Chloro-3-oxo-1-phenylbutan-2-yl)isoindoline-1,3-dione (Cl-K), 140.18, 140.23, 140.46, 141.69, 141.81, 141.95, 142.50, 142.75,
70% yield, m.p. 118–119 1C (lit. m.p. 117–118 1C).23 1H NMR 142.85, 142.96, 143.19, 143.32, 143.41, 143.73, 143.81, 144.03,
spectrum (acetone-d6, d ppm, J Hz): 3.39 (dd, 1Ha, CH2, 144.28, 144.46, 144.51, 144.58, 144.76, 144.83, 144.91, 145.06,
3
2
3
2J 14.2, J 5.1), 3.62 (m, 1Hb, CH2, J 14.2, J 10.9), 4.22 (s, 2H, 145.30, 145.51, 145.64, 145.69, 145.85, 146.06, 146.17, 146.31,
3
3
Cl–CH2), 5.29 (dd, 1H, CH, J 5.1, J 10.9), 7.13 (m, 5H, C6H5), 146.46, 146.65, 147.01, 147.09, 147.25, 147.32, 167.45, 167.66,
7.69–7.80 (m, 4H, C6H4).13C NMR spectrum: 33.85, 46.14, 58.00, 167.71, 167.76, 167.80, 167.86, 195.52. UV-Vis (CH2Cl2, nm): 256,
123.69, 127.06, 128.63, 128.96, 131.31, 134.43, 136.07, 167.43, 485, 693. FT-IR (m.o.): 2922, 2850, 1775, 1715, 1383, 1336, 797,
197.13. Found, %:
C 65.97, H 4.30, N
4.26, Cl 10.75. 719. C96H26N2O6 MS (MALDI-TOF), m/z: 1302.375 (M+) (calc.,
C18H14ClNO3. Calculated, %: C 65.96, H 4.31, N 4.27, Cl 10.82. m/z: 1302.179).
2-(4-Bromo-3-oxo-1-phenylbutan-2-yl)isoindoline-1,3-dione (Br-K),
75% yield, m.p. 144–145 1C.22 1H NMR spectrum (acetone-d,6 d
Chromatographic research techniques
2
3
ppm, J Hz): 3.38 (dd, 1Ha, CH2, J 14.2, J 5.1), 3.59 (m, 1Hb,
CH2, 2J 14.2, 3J 11), 4.01 (s, 2H, Br–CH2), 5.38 (dd, 1H, CH, 3J 5.1,
3J 11), 7.11 (m, 5H, C6H5), 7.68–7.81 (m, 4H, C6H4). 13C NMR
spectrum: 31.36, 33.97, 57.69, 123.68, 127.03, 128.63, 128.96,
131.33, 134.45, 136.14, 167.44, 197.86. Found, %: C 58.13,
H 3.76, N 3.75, Br 21.42. C18H14BrNO3. Calculated, %: C 58.08,
H 3.79, N 3.76, Br 21.47.
Procedure for the preparation of MMF and DMF-functionalized
methanofullerenes.
Chloroketone Cl-K 0.045 g (0.138 mmol) (or bromoketone
Br-K 0.051 g (0.138 mmol)) and 0.053 g (0.414 mmol) DBU was
added to a solution of 0.1 g (0.138 mmol) of C60 fullerene in
HPLC system. In this work, we used the method of reverse
phase chromatography to identify fullerene-containing compounds
by HPLC. Kinetic studies were carried out using a SHIMADZU LC-20
Prominence liquid chromatograph (Japan) in the basic con-
figuration without a diode array detector because the measurements
were carried out at specific wavelengths with the maximum
simplified analysis and a variable wavelength programmable
UV/Vis detector. The results were processed using the Shimadzu
chromatographic equipment LabSolutions application package.
Chromatographic conditions
35 ml toluene. The reaction mixture was stirred for 15 to 20 min The HPLC experimental conditions were optimized on a SUPELCO-
at room temperature (monitored by TLC). Then, the mixture SIL C18 column (150 mm ꢀ 4.6 mm internal diameter, 5 mm
was washed with 20 ml of 5% HCl solution, the organic phase particle size) with a pre-column.24 The optimum mobile phase was
was separated and dried with MgSO4, the solvent was removed prepared by mixing toluene and acetonitrile in the ratio of 1 : 1 (v/v).
in vacuum and the residue was separated by silica gel column The elution mode was isocratic. A wavelength of 330 nm of the
chromatography (toluene/petroleum ether, 5 : 1, v/v).
diode-array detector was chosen because C60 and the adducts of
2-{1-Benzyl-2-[cyclopropano[1f,2f ][60]fullerene]-2-oxoethyl}-1H- cycloaddition have sufficient absorption at this wavelength. The
isoindole-1,3(2H)-dione (MMF), dark brown powder, yield 50%. flow rate used was 1.0 ml minꢁ1. The injection volume was 20 ml
1H NMR spectrum (CDCl3, d ppm, J Hz): 3.71 (dd, 1Ha, CH2, and the temperature of the column and flow cell spectrophoto-
3
2
3
2J 14.3, J 5.5), 3.99 (dd, 1Hb, CH2, J 14.3, J 10.5), 4.98 (s, 1H, meter was 35 1C. The retention times were 1.81 min for the DMF
CH); 5.71 (dd, 1H, CH, 3J 5.5, 3J 10.5); 7.11 (m, 5H, C6H5), 7.73 and adduct, 2.16 min for the MMF adduct and 4.98 min for fullerene
7.89 (dm, 4H, C6H4). 13C NMR spectrum: 34.12, 42.79, 60.89, C60. The total run time of the system was about 7 min. For
71.23, 71.82, 123.90, 125.32, 127.22, 128.25, 128.87, 129.06, the tested products, the UV spectra and the retention times
128.12, 131.46, 134.63, 136.34, 136.50, 137.88, 140.21, 140.36, were used as identification parameters (see ESI,† page S12,
140.96, 141.10, 141.15, 142.06, 142.15, 142.25, 142.27, 142.70, paragraph 2.1).
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
Paper
NJC
Preparation of standard solution
structures presented correspond to the minima on the potential
energy surface. The latter fact was confirmed by obtaining a set
of only positive vibration frequencies by Hessian calculations.
All reported reaction and activation energies at 298 K are
compared to the electronic energies to allow direct comparison
between theoretical and experimental data.
To quantitatively determine the studied substances, we proposed
an absolute calibration method. The samples were weighed on a
Sartiorius M2P balance (sensitivity 10ꢁ6 g). Calibration solutions
of the compounds were prepared by direct dissolution of full-
erene and the methano derivatives in a toluene/acetonitrile
mixture (1 : 1, v/v). The concentration of the solutions ranged
from 10ꢁ6 to 10ꢁ4 molꢁ1; aliquots were taken using BiohitProline
series micropipettes (Finland). The mobile phase was filtered
using a 0.45 mm microporous filter. Good separation of the test
substances was observed, and the resulting peaks had regular
and symmetrical shapes under the developed conditions (see
ESI,† pages S13 and S14, paragraph 2.2).
Results and discussion
Methanofullerene synthesis
Because fullerene C60 possesses 30 equally active double bonds
at the corresponding [6,6]-ring transitions, it is natural to
expect numerous occurrences of C60 cyclopropanation of the
nucleus rather than a single one. The interaction can be carried
out selectively towards the formation of either a monoadduct or
several adducts with a well-defined addition scheme. Halogenated
substrates with an activated halogen–methine function capable of
forming stabilized carbanions are the key C60 cyclopropanating
agents under the ‘‘classical’’ conditions of the Bingel reaction in
methanofullerene synthesis, and intramolecular cyclization is
provided by the good leaving group halide (mainly Clꢁ).19
Fullerenes are powerful antioxidants (in terms of antioxidant
effects, they are superior to known analogues by more than two
orders of magnitude),36–40 and the presence of an N-substituted
terpene fragment in an amino acid residue reliably provides an
anti-inflammatory effect and facilitates the penetration of the
drug into the intracellular space; therefore, joining them into
one molecule will enable the advantageous combination of their
properties, and new qualities will be expected to appear.
Another reason to develop such compounds is the worldwide
demand in medicine for multifunctional drugs (such as those
combining antioxidant and antiviral properties).
Preparation and kinetic study of sample solutions
The standard solutions of the starting reagents were prepared
by dissolving 108 mg C60, 49 mg Cl-K (or 56 mg Br-K) and 23 mg
DBU in 50 ml of toluene in four separate 50 ml volumetric
flasks. Final volumes were adjusted with toluene to prepare
3 mg mlꢁ1 standard stock solutions for all reagents. Using a
volumetric pipette, certain molar ratios of these solutions were
transferred to a glass reactor equipped with a magnetic stirrer
and completed to a volume of 2 ml using toluene. The flasks
were placed in a thermostatic oven at different temperatures
(294, 308, 323, and 338 K) for different time intervals (1–6 h). At
specified time intervals, separate 40 ml aliquots of the reaction
mixture were transferred into a series of calibrated separate
stoppered conical flasks and dissolved in 2 ml of the mobile
phase. Then, the final test solution was washed with 5% HCl
solution, dried with MgSO4, filtered using a 0.45 mm micro-
porous filter and subjected to the chromatographic conditions.
The concentrations of the compounds were evaluated on the
basis of the ratios of their peak areas to those of the standard
solutions. To evaluate the reproducibility, each experiment was
performed twice.
To solve the stated problems, unexplored methanofullerenes
as target products based on cyclopropanation agents were
recently synthesized according to Scheme 1.
Phthalic anhydride and b-phenyl-a-alanine, taken in a 1 : 1
molar ratio, were directly fused to acid I, which was converted
Theoretical modeling of the kinetic process
The kinetic experiment was processed to calculate the reaction
rate constants using the MAPLE 16 computer mathematics
system.25–27
Computational details
Quantum-chemical calculations were carried out using the
Gaussian 09 software package.28 The calculation results were
visualized using the Chemcraft 1.8029 and GaussView 5.030 pro-
grams. The calculations were performed using the 6-31G(d) basis
set,31,32 which is a standard split-valence basis with addition of a
set of d-polarization functions for non-hydrogen atoms. Electron
correlation was taken into account at the DFT density functional
theory level using the B3LYP hybrid functional.33–35 Full geometry
optimizations of all the structures of interest were performed, and
frequency calculations were performed. Solvent effects were
estimated with single-point calculations on the gas phase
optimized structures using the polarizable continuous solvation
model (PCM) and considering toluene as the solvent. All the
Scheme 1 Synthesis of halogenmethyl ketones from diazocompounds.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
NJC
Paper
Fig. 1 Time-dependent HPLC charts (SUPELCOSIL, toluene/acetonitrile =
1 : 1, 1 ml minꢁ1) for the Bingel reaction of Cl-K with C60 in toluene at
330 nm (with the same charts for the reaction of Br-K with C60 in toluene).
Scheme 2 Synthesis and chemical structures of the mono-adduct MMF
and bis-adduct DMF.
At the same time, the process was monitored by observing
the initial fullerene consumption and accumulation of mono-
and bis-adducts at a quantitative level. The UV spectra of
the determined substances and retention times were used as
identification parameters.
Typical chromatograms of the reaction mixture obtained by
monitoring the reaction of chloromethyl ketone Cl-K (or bromo-
methyl ketone Br-K) and C60 fullerene under DBU in toluene
during the process are shown in Fig. 1.
to acid chloride II with thionyl chloride. The latter, without any
isolation from the reaction mixture, was treated with a diazo-
methane solution in methylene chloride to obtain diazoketone
III; its reaction with aqueous HCl led to chloromethyl ketone
Cl-K and its reaction with aqueous HBr led to bromomethyl
ketone Br-K.
Further, methanofullerenes with a phthalimide block of
MMF and DMF adducts were isolated by Bingel reactions of
halogenmethyl ketones and C60 fullerene under DBU (toluene).
Their structures were established by modern spectral analysis
methods (Scheme 2) (see ESI,† pages S2–S8, paragraph 1). We
note that halogenmethyl ketones as cyclopropanation agents
were first studied under the Bingel process.41
A decrease in the chromatographic peak of the initial fullerene
C60 content and increases in the MMF and DMF chromatographic
peaks are clearly observed.
Chromatographic studies of various molar ratios of fullerene
and the chloromethyl ketone reaction mixture (0.5 : 1; 1 : 1; 1 : 2;
and 1 : 4 mol molꢁ1, respectively) allowed us to obtain kinetic
curves of the consumption of the initial fullerene and accumulation
of the MMF and DMF adducts (Fig. 2).
Kinetics of the Bingel reaction of fullerene C60 and
halogenmethyl ketones
Fig. 2 clearly demonstrates a characteristic increase in the
consumption rate of the initial C60 in the reaction mixture and
After analyzing the substances by HPLC, a quantification
method was developed.
Initially, the absolute calibration method (or external standard
method) was applied to quantify the studied substances. The area
of the chromatographic peak was used as an analyte content
parameter in the sample. Standard C60 and its MMF and DMF
derivative solutions in a toluene and acetonitrile mixture were
separately prepared (1 : 1, v/v) to construct a calibration graph. A
linear dependence between the solution concentration and the
chromatographic peak area occurred over the entire range of the
selected concentrations (see ESI,† pages S13 and S14, paragraph 2.2).
The kinetics of the conversion of C60 to methanofullerenes
was studied under homogeneous catalysis conditions using a
reverse-phase HPLC method for which the best chromatographic
conditions were preliminarily selected (see ESI,† pages S15–S16,
paragraph 2.3).
Thus, based on the data of optimizing the eluent composition
(acetonitrile/toluene, 1 : 1, v/v), gradient chromatography conditions
were revealed under the isocratic mode, where the C60 retention
times of the MMF and DMF functionalized methanofullerenes were
4.98, 2.16 and 1.81 min, respectively. This enabled us to increase the
retention substance time to improve their separation and to exclude
possible overlapping of the peaks of the investigated substances.
Fig. 2 Kinetic curves for the consumption of fullerene C60
tion of monoadduct MMF ( ) and diadduct DMF ( ) over the reaction time with
(a) 0.5, (b) 1.0, (c) 2.0 and (d) 4.0 equiv. of Cl-K in the initial mixture.
( ) and accumula-
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
Paper
NJC
the accumulation of the mono- and bis-functionalization products ratio is rarely used. Due to the high-speed cyclopropanation
with increasing molar fraction of chloromethyl ketone in the initial process of fullerene and, accordingly, the difficulties in con-
mixture. It is noteworthy that the intersection of the C60 con- trolling the reaction kinetics, a 1 : 4 molar ratio was not used.
sumption curves and MMF accumulation curves always occurs This ratio is advantageous only for the preparation of C60
at the fullerene conversion degree of B55–56%. Similar conversion polyadducts, which is beyond the scope of this study.
curves were observed during the reaction of bromomethyl ketone
and fullerene.
The table data indicate that as the process temperature
increases, the mono-adduct yield control becomes more difficult;
Various experimental design methods have been used to however, at the same time, the bis-adduct yield becomes optimal.
determine essential parameters for the studied reaction and its The reaction time considerably affects the C60 conversion. The
optimization.
Firstly, Table 1 reveals the temperature, process time, and by the kinetic curves (see ESI,† page S17, paragraph 3).
molar ratios of the initial reagents as key factors in the conversion For example, a 2-fold increase in the substrate concentration
halogenmethyl ketone concentration effect can also be estimated
of C60 to MMF :DMF. For further study of the reaction, it seemed with respect to fullerene leads to a proportional increase in the
important to compare the relative contributions of these factors reaction rate. It seems that the optimal Bingel conditions for
and identify any possible multifactorial interactions.
mono-adduct synthesis are a twofold excess of halogenmethyl
As a rule, the process of cyclopropanation of C60 according ketone with respect to C60 and 294 K (room temperature)
to the Bingel method is carried out at room temperature. (Table 1, entries 7–9), whereas an equimolar amount of the
Therefore, we tested four variants of the molar ratio of the substrate and 338 K temperature are optimal for production of
starting reagents at the indicated temperature. Most studies the bis-adduct (Table 1, entries 21 and 22).
use an excess of cyclopropanation agent in this process.1,18,42,43
Secondly, to determine the optimal ratio of starting reagents
In this regard, at elevated temperatures, the kinetics of the for predicting the mono-functionalized adduct yield, the kinetic
reaction was considered in two proportions. One proportion is experiment data were converted into calibration curves by conver-
equimolar, to determine the activation parameters, and the sion of the C60 coordinates vs. MMF yield (DMF yield or MMF :DMF
other is a twofold excess of halogenmethyl ketone, as an ratio) (Fig. 3). As can be seen, the results are almost identical.
example of the most frequently used ratio for the starting Therefore, varying the ratio of starting reagents leads to rate changes
reagents during fullerene functionalization. The C60 : substrate only, and the yields of both mono- and bis-functionalized adducts
ratio of 1.0 : 0.5 mol molꢁ1 was not considered because this remain the same. The maximum possible MMF content is B55%
with these cyclopropane agents (Fig. 3a). Further processing helps to
reduce the mono-functionalized adduct concentration as a result of
its quantitative transition to bis- and polyadducts. A 28% yield of the
Table 1 Results for the reactions between fullerene and halogenmethyl
adduct DMF is achieved by 83% fullerene conversion (Fig. 3b).
A similar predictive model in the exponential form of C60 conversion
vs. the MMF :DMF ratio calibration curves is also traced in the
publication.8 The construction of such model dependences from a
ketones in toluene
Product ratio (%)
Time
(min)
Entry
Equiv. Cl-K/Br-Ka
T (K)
MMF
DMF
1
2
3
4
5
6
7
8
0.5
294
30
60
120
30
60
120
30
60
120
30
60
15.0/13.5
24.1/22.6
35.4/29.8
26.5/24.3
32.1/31.7
40.9/38.6
32.0/29.9
42.9/40.1
50.4/47.7
46.3/46.0
54.0/52.1
56.4/49.9
1.5/1.4
2.5/2.1
5.3/4.6
2.0/2.0
4.1/3.9
7.5/6.6
4.2/3.7
8.0/7.5
13.7/12.4
8.8/8.4
15.8/14.2
24.2/22.3
1
2.0
4.0
9
10
11
12
120
13
14
15
16
1.0
2.0
308
323
338
30
60
30
60
27.5/25.9
46.6/45.0
46.7/45.8
48.2/48.0
16.2/15.9
24.9/25.3
25.3/25.1
23.2/23.1
17
18
19
20
1.0
2.0
30
60
30
60
50.5/50.3
46.3/47.0
52.3/52.0
44.3/44.9
24.6/25.4
22.6/23.1
21.4/22.3
18.2/19.0
21
22
23
24
1.0
2.0
30
60
30
60
54.1/54.1
53.1/54.4
42.4/43.0
41.5/42.1
27.6/26.2
27.1/27.4
25.4/26.0
23.8/24.2
Fig. 3 Calibration curves relating to (a) C60 conversion vs. MMF yield;
(b) C60 conversion vs. DMF yield; (c) C60 conversion vs. MMF : DMF ratio.
Equiv. of Cl-K in the initial mixture: ( ) 0.5; ( ) 1.0; ( ) 2.0; ( ) 4.0.
a
DBU equiv. amount to halogenmethyl ketones.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
NJC
Paper
the initial and current C60, MMF, and DMF concentrations at
various points in time. The kinetic curves were reproduced by
mathematical modeling. As confirmed in the developed program,
the experiment interpretation accuracy allows us to find the rate
constants of the experimental data at any temperature (see Table 2).
To find the activation parameters of C60 cyclopropanation
according to the Bingel reaction, chromatographic studies were
performed in the temperature range of 294 to 338 K. Processing of
experimental data in the coordinates of the Arrhenius equation
allowed us to determine the activation parameters of the Bingel
reaction (see ESI,† page S18, paragraph 4).
Scheme 3 Cyclopropanation of fullerene with the halogen-containing
substrate and the non-nucleophilic base.
The activation energies for the first stage of Cl-K and Br-K
cyclopropanation were 15.9 and 16.1 kcal molꢁ1, respectively.
The presented experimental data indicate that when choosing
between chloromethyl ketone and bromomethyl ketone as the
cyclopropanation agent, it is preferable to use the former due to
its higher efficiency.
given set of reaction conditions allows one to reliably predict the
yields of both mono- and bis-functionalized methanofullerenes.
Based on the set of experiment kinetic curves, the inverse
kinetic problem as a validation of the proposed reaction
mechanism has been solved. It consists of the initial deprotonation
of substrate B by the strong base C (in our case, DBU) to form the
reactive nucleophile Bꢁ, which attacks the electron-deficient double
fullerene bond (A). The resulting intermediate TS carbanion is
transformed into the methano derivative D as a result of intra-
molecular SN2 substitution of the halogen atom (Scheme 3).18
In order to determine the rate constants of the Bingel
cyclopropanation process, the process kinetics was theoretically
simulated using the MAPLE 16 computer mathematics system.
Two main kinetic stages of the process observed by HPLC were
considered:
Computational studies on the Bingel reaction of C60 with
halogenmethyl ketones
Density functional theory, especially in recent years, has been
demonstrated as a powerful method for studying detailed reaction
mechanisms as well as for predicting stereo and chemoselectivity
in organic reactions. In the scientific literature, only the cyclo-
addition of bromomalonates to afford higher endohedral full-
erenes has been considered while studying the theoretical aspects
of Bingel methanofullerene synthesis using DFT.12–17
In our work, we performed a quantum chemical simulation
of the C60 cyclopropanation reaction containing the halogen-
methyl ketone phthalimide block by the Bingel method. First,
we optimized the structures of all cyclopropanation participants
in the first stage leading to the MMF adducts.
k1
ƒ!
A þ B þ C
D
E
k2
D þ B þ C
ƒ!
When studying the kinetic aspects of complex processes,
In this first step, C60 (A) in situ reacts with the halogenmethyl various stationary points in the reaction are calculated. Initially,
ketone (B) under DBU (C) to form a mono-functionalized MMF an intermediate compound is formed with halogenmethyl
(D). Further, adding (B) (the second stage) under (C) forms a ketone bound to the C60 core, which gives a negative charge to
discontinuous DMF (E) with the corresponding reaction rate the entire system. The halogen anion is then released through a
constant.
transition state during which a second bond is formed between
The kinetic experiment shows that the differences in the the carbon halogenmethyl ketone and fullerene atom. The
concentration of the monitored participants of the spending speed-determining stage of the Bingel process is the conversion
and accumulation reactions with time are less than 1% (Fig. 1). of the negatively charged intermediates into products (the
Therefore, further steps of the process that generate various formation of the reactant complex is a barrier-free process).12
cycloaddition polyadducts are of no interest and are not con-
sidered in this research.
The general optimized characteristics of the intermediates
(MMF-Int(Cl) or MMF-Int(Br)) and transition complexes (MMF-
The reaction rate constants of mono-k1 and bis-cycloaddition k2 TS(Cl) or MMF-TS(Br)) are shown in Fig. 4 (a list of Cartesian
were calculated using a MAPLE program (see ESI,† pages S19–S21, coordinates for the optimized structures is presented in ESI,†
paragraph 5) while taking into account the experimental values of pages S22–S27, paragraph 6).
Table 2 Calculated rate constants (10ꢁ5 l2 molꢁ2 minꢁ1) for the reaction of C60 with halogenmethyl ketones in the presence of DBU
Substratea
Reaction temperature (K)
294
308
323
338
Cl-K
k1
k2
k1
k2
1.21 ꢂ 0.02
0.50 ꢂ 0.02
1.12 ꢂ 0.03
0.49 ꢂ 0.02
3.19 ꢂ 0.04
1.32 ꢂ 0.04
3.11 ꢂ 0.04
1.29 ꢂ 0.03
13.84 ꢂ 0.05
5.65 ꢂ 0.05
11.12 ꢂ 0.04
4.63 ꢂ 0.04
40.81 ꢂ 0.06
15.69 ꢂ 0.06
38.13 ꢂ 0.06
15.56 ꢂ 0.05
Br-K
a
Substrate equiv. amount to C60
.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
Paper
NJC
thermal effects of the conversions under discussion compared to
the system of the starting reagents are ꢁ24.1 and ꢁ24.4 kcal molꢁ1
,
respectively; therefore, the Bingel process proceeds under thermo-
dynamic control.
Conclusions
The selective synthesis of new methanofullerene derivatives by
the Bingel reaction was carried out using halogenmethyl ketones as
cyclopropanating agents.
The experimental study of the kinetics of the reaction of
fullerene with a cyclopropanation agent containing a phthalimide
block was carried out using HPLC, with quantitative monitoring of
the progress of the process by the consumption of C60 and the
accumulation of mono- and di-adducts. The construction of calibra-
tion curves in various coordinates revealed that varying the ratio of
the starting reagents does not affect the yield of either mono- or bis-
functionalized adducts; regardless of the nature of the cyclopropa-
nation agent, these yields are 55% and 28%, respectively.
Quantum-chemical modeling of this cyclopropanation reaction
has been carried out under the DFT method. The activation
parameters of the theoretical and experimental approaches correlate
with each other. It was revealed that the chloromethyl ketone is
preferable to its bromo analogue in C60 core functionalization; also,
a twofold excess of the substrate relative to fullerene is the optimal
ratio for the synthesis of mono-adducts.
Fig. 4 Spatial structures of the MMF-Int(Cl) and MMF-Int(Br) intermediates
and the transition complexes MMF-TS(Cl), MMF-TS(Br) in the cyclo-
propanation reactions optimized by the B3LYP/6-31G (d) method: bond
lengths in Å, angles in degrees.
As a result, the proposed HPLC method can be adopted for
quantitative quality control and routine analysis of fullerene con-
version to mono- and polyfunctionalized methanofullerenes.
Fig. 5 Energy profile of the first stage of fullerene C60 cyclopropanation
with chloromethyl ketone Cl-K (blue line) and bromomethyl ketone Br-K
(red line) with MMF adduct formation (B3LYP/6-31G(d) approximation,
toluene solvent, 298 K).
Conflicts of interest
There are no conflicts to declare.
As can be seen, the distance between the central carbon
atom of the halogenmethyl ketone and the halogen in the TS
structure is greater than in the intermediate, which indicates
the release of the halogen anion. Moreover, it is important to
note the change in the hybridization of the central carbon atom
of the halogenmethyl ketone group from sp2 in the intermediate
to sp3 in TS.
Acknowledgements
The work was financially supported by the Ministry of Education
and Science of the Russian Federation within the framework of
the basic part of the state task in the field of scientific activity
(project no. 4.9143.2017/8.9) and the framework of state assign-
ment (project no. AAAA-A19-119020890014-7).
Fig. 5 reveals the interaction energy profile in toluene,
obtained by analyzing the potential energy surface and relaxing
scanning of the reaction coordinates at the DFT level using the
B3LYP/6-31G(d) method.
References
1 C. Bingel, Chem. Ber., 1993, 126, 1957–1959.
2 Y. N. Biglova and A. G. Mustafin, RSC Adv., 2019, 9, 22428–22498.
3 H. Ueno, H. Kawakami, K. Nakagawa, H. Okada, N. Ikuma,
S. Aoyagi, K. Kokubo, Y. Matsuo and T. Oshima, J. Am.
Chem. Soc., 2014, 136, 11162–11167.
4 N. Ikuma, Y. Susami and T. Oshima, Eur. J. Org. Chem.,
2011, 6452–6458.
5 T. Oshima, T. Mikie, N. Ikuma and H. Yakuma, Org. Biomol.
Chem., 2012, 10, 1730–1734.
According to the energy diagram, both halogenmethyl
ketones form stable intermediates with fullerene because their
total calculated energies have lower values than the starting
compounds. It is noteworthy that chloromethyl ketone Cl-K
during C60 cyclopropanation obtains a more effective stabilization
than a similar pre-reaction complex, bromomethyl Br-K. Further
transformation of the intermediate into a mono-adduct through a
transition state is characterized by energy barriers of 19.7 and
22.0 kcal molꢁ1 for MMF-TS(Cl) and MMF-TS(Br), respectively,
which is generally consistent with the experimental data. The overall
6 T. Mikie, H. Asahara, K. Nagao, N. Ikuma, K. Kokubo and
T. Oshima, Org. Lett., 2011, 13, 4244–4247.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020

View Article Online
NJC
Paper
7 T. Oshima, H. Kitamura, T. Higashi, K. Kokubo and
N. Seike, J. Org. Chem., 2006, 71, 2995–3000.
8 A. Bianco, M. Maggini, M. Nogarole and G. Scorrano, Eur.
J. Org. Chem., 2006, 2934–2941.
9 J. P. Barham, S. Tanaka, E. Koyama, N. Ohneda, T. Okamoto,
H. Odajima, J. Sugiyama and Y. Norikane, J. Org. Chem.,
2018, 83, 4348–4354.
10 N. Ikuma, Y. Susami and T. Oshima, Org. Biomol. Chem.,
2010, 8, 1394–1398.
11 S. Silvestrini, D. DalleNogare, T. Carofiglio, E. Menna,
P. Canu and M. Maggini, Eur. J. Org. Chem., 2011, 5571–5576.
12 C.-X. Cui and Y.-J. Liu, J. Phys. Org. Chem., 2014, 27, 823–832.
13 N. Alegret, A. R. Fortea and J. M. Poblet, Chem. – Eur. J.,
2013, 19, 5061–5069.
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin,
V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari,
A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski,
G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D.
¨
Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and
D. J. Fox, Gaussian 09, Revision C.01, Gaussian Inc., Wallingford
CT, 2009.
`
´
14 M. Garcia Borras, M. R. Ceron, S. Osuna, M. Izquierdo, 29 G. A. Zhurko, Chemcraft 1.80 (build 536) – graphical program
`
J. M. Luis, L. Echegoyen and M. Sola, Angew. Chem., Int. Ed.,
for visualization of quantum chemistry computations, https://
2016, 55, 2374–2377.
chemcraftprog.com.
´
`
15 J. P. Martınez, M. Garcia Borras, S. Osuna, J. Poater, F. M. 30 R. D. Dennington II, T. A. Keith and J. M. Millam, GaussView
`
Bickelhaupt and M. Sola, Chem. – Eur. J., 2016, 22, 5953–5962.
16 N. Alegret, P. Salvado, A. R. Fortea and J. M. Poblet, J. Org. 31 W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys.,
Chem., 2013, 78, 9986–9990. 1972, 56, 2257–2261.
17 S. Osuna, R. Valencia, A. Rodrıguez Fortea, M. Swart, M. Sola 32 P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973, 28,
and J. M. Poblet, Chem. – Eur. J., 2012, 18, 8944–8956. 213–222.
18 Q.-Z. Li, J.-J. Zheng and X. Zhao, J. Phys. Chem. C, 2015, 119, 33 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
5.0, Wallingford CT, 2009.
´
`
26196–26201.
34 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785–789.
35 B. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38,
3098–3100.
36 M. Ehrich, R. V. Tassell, Y. Li, Z. Zhou and C. L. Kepley,
Toxicol. Cell. in Vitro, 2011, 25, 301–307.
19 A. Hirsch, The chemistry of the fullerenes, Wiley-VCH Verlag
GmbH & Co. KGaA, Weinheim, 2002, p. 223.
20 J. P. Greenstein and M. Winitz, Chemistry of the Amino Acids,
John Wiley & Sons, New York, 1961, p. 763.
¨
21 D. M. Shendage, R. Frohlich and G. Haufe, Org. Lett., 2004,
6, 3675–3678.
37 Z. Zhou, R. P. Lenk, A. Dellinger, S. R. Wilson, R. Sadler and
C. L. Kepley, Bioconjugate Chem., 2010, 21, 1656–1661.
38 V. Sergeeva, O. Kraevaya, E. Ershova, L. Kameneva,
E. Malinovskaya, O. Dolgikh, M. Konkova, I. Voronov,
A. Zhilenkov, N. Veiko, P. Troshin, S. Kutsev and S. Kostyuk,
Oxid. Med. Cell. Longevity, 2019, 4398695.
22 M. Kolb, B. Neises and F. Gerhart, Liebigs Ann. Chem., 1990,
1–6.
23 C. L. Liotta, P. Pollet, K. Kitagawa, W. Dubay and J. Stringer,
Reduction of aldehydes and ketones, WO 2011060302, 2010.
24 K. Jinno, Y. Sato, H. Nagashima and K. Itoh, J. Microcolumn
Sep., 1998, 10, 79–88.
¨
39 E. Oberdorster, Environ. Health Perspect., 2004, 112, 1058–1062.
25 A. G. Makeev, M. M. Slinko and D. Luss, Appl. Catal., A, 40 N. Gharbi, M. Pressac, M. Hadchouel, H. Szwarc, S. R.
2019, 571, 127–136.
Wilson and F. Moussa, Nano Lett., 2005, 5, 2578–2585.
41 Y. N. Biglova, R. N. Malikova, S. F. Petrova, S. P. Ivanov,
I. M. Sakhautdinov and A. G. Mustafin, Int. J. Chem. Kinet.,
2019, 51, 311–320.
26 Y. Wang and S. Liu, Energies, 2014, 7, 8411–8426.
27 P. Jouvet, P. Hubert, J. M. Saudubray, D. Rabier and
N. K. Man, Pediatr. Res., 2005, 58, 278–282.
28 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, 42 D. Armspach, E. C. Constable, F. Diederich, C. E. Housecroff
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, and J.-F. Nierengarten, Chem. Commun., 1996, 2009–2010.
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, 43 X. Camps, H. Schonberger and A. Hirsch, Chem. – Eur. J.,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, 1997, 3, 561–567.
¨
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2020